Could you please tell me how many CAN interfaces are there in STM32H723

[zzzaizz]

Could you please tell me how many CAN interfaces are there in STM32H723? I asked the AI, and the AI said there are 2. But when I configured it using Cubemx, I found there are 3. Help me, please!

[Ozone]

Don't rely on "Artificial Interference".

Here is a comparison table provided by ST : https://www.st.com/en/microcontrollers-microprocessors/stm32h723-733/products.html

All of the MCU variants in this table list 3 CAN interfaces. But I would check with the datasheet of the exact package variant you intending to use.

[Andrew Neil · Accepted solution]

As @Ozone said, don't rely on  AI.

You must always verify what any AI tells you - they are known to make errors, and to make stuff up ("hallucinate")!

The definitive answer will always be found in the datasheet for the chip in question.

For best results, always go to the manufacturer's own website - third-party sites are not necessarily up-to-date.

This is not specific to STM32 or even  ST - this applies to any electronic component.

PS:

---

@zzzaizz wrote:

I asked the AI

---

Which AI did you ask?

Note that ST has its own AI assistant - called "Sidekick" - which is specifically trained on STM32 documentation.

Even so, it does sometimes get it wrong.

But  it does give you links to datasheets, etc - so you can verify what it says.

PPS:

Yes, I know that some people frown on the use of "hallucination" for what AI does (hence the quotes), but you know what I mean!

[Ozone]

> Yes, I know that some people frown on the use of "hallucination" for what AI does (hence the quotes), but you know what I mean!

Not "frown" in my case, but I view this as euphemism.
I'm old enough to remember the inception of this idea, called "Expert Systems" back then. And basically is today what it was back then - a bunch of algorithms combined with a (dynamic) database.
This works fine in technical fields with a strictly defined and constrained language, and strictly encoded axioms (laws of physics).

The disaster area starts when this idea is rolled out to pseudo.scientific fields, especially politics. Due to encoded political biases one can find an "AI" contradicting itself several times within a single conversation, or have quasi-emotional meltdowns.

By the way, I use the term "pseudo science" according the definition of Karl Popper, i.e. for fields without falsifiable hypotheses, and without repeatable experiments. And there are more of those than one might think ...

[zzzaizz]

Thank you. I will make good use of artificial intelligence and not be so hasty anymore!

[zzzaizz]

Thank you.